Region of Interest Definition in Cardiac Imaging

1. A method for cardiac imaging, comprising: acquiring functional imaging data of a subject, the functional imaging data including at least a myocardium; determining a region of interest (ROI) encompassing at most the myocardium from the acquired functional imaging data, determining the ROI including: generating projection image representations from the functional image data, thresholding the projection image representation based on a histogram of a selected representative slice, determining a localized cardiac region within the thresholded data according to at least one of a horizontal ray casting profile and anatomical heuristics, and determining extents of the localized cardiac region according to image based and heuristics based cleaning techniques; and estimating diagnostic parameters of the myocardium based on the determined ROI.

2. The method according to claim 1 , wherein the estimated diagnostic parameters include at least one of myocardial blood flow, regional myocardial blood flow, flow reserve, and ejection fraction.

3. The method according to claim 1 , wherein thresholding the acquired functional data includes: generating a histogram of the selected representative slice; determining a maximum frequency and its index from the histogram; searching for a local maximum frequency and its index based on preselected conditions on the basis of the index of the maximum frequency; and iteratively determining a thresholding factor based on at least one of an index and frequency of the local maximum frequency.

4. The method according to claim 3 , wherein thresholding the projection image representation further includes: determining a benchmark line from a ratio of the local maximum frequency to the determined thresholding factor; locating a first benchmark frequency and its index on the histogram, the benchmark frequency being a first frequency below the benchmark line; locating a last benchmark frequency and its index, the last benchmark frequency having a frequency between the first benchmark frequency and the benchmark line and an index between three to fifty of the first benchmark point index; locating the thresholding frequency and its index, the threshold frequency being lowest frequency between the first and last benchmark frequencies; and thresholding the projection image representation such that the frequencies with indices greater than or equal to the second benchmark index are preserved.

5. The method according to claim 4 , wherein determining a localized cardiac region within the thresholded data include: generating the ray casting profile of the thresholded projection image representation from horizontally casted rays; determining ray which are connected according to connectivity criteria; and determining the localized cardiac region according to anatomical heuristics to the connected rays.

6. The method according to claim 5 , wherein the anatomical heuristics defining a localized cardiac region according to approximate cardiac dimensions and anatomical position based on non-zero pixel counts of casted rays.

7. The method according to claim 5 , wherein in response to the determined localized cardiac region failing, updating the thresholding factor and thresholding the projection image representation according to an updated benchmark line and/or frequencies.

8. The method according to claim 5 , wherein determining the extents of the localized cardiac region includes: masking the determined localized cardiac region; cleaning the masked localized cardiac region using image based cleaning techniques; determining extents of cleaned localized cardiac region; adjusting extents of cleaned localized cardiac region according to vertical ray casting based heuristics; and estimating a bounding box for the adjusted extents.

9. The method according to claim 1 , further including: reconstructing the functional imaging data which correspond to the determined extents of the localized cardiac region.

10. A method for cardiac imaging, comprising: acquiring functional imaging data of a subject, the functional imaging data including at least a myocardium; determining a region of interest (ROI) encompassing at most the myocardium from the acquired functional imaging data, determining the ROI including: reconstructing the function imaging data into a volume image representation; thresholding transverse slices of the volume image representation according to a maximum pixel intensity of each corresponding slice; determining a cluster of the thresholded slices which correspond to a left ventricle; cleaning the determined cluster of slices according to a predetermined Manhattan distance between a left ventricle centroid in a central slice to the left ventricle centroid in the remaining clustered slices; and isolating the left ventricle in the cleaned slices according to cardiac angles of a septum of the myocardium; and estimating diagnostic parameters of the myocardium based on the determined ROI.

11. The method according to claim 10 , further including: determining the azimuth and elevation angle of the isolated left ventricle; determining a short axis, a vertical long axis, and a horizontal long axis of the left ventricle according to the determined azimuth and elevation angle; and determining reorientation angles of the left ventricle according to the determined short axis, vertical long axis, and horizontal long axis.

12. The method according to claim 10 , wherein thresholding the transverse slices includes: determining a maximum pixel value for each transverse slice; and preserving pixels within each slice with intensities above a ratio of the corresponding maximum pixel value to a preselected thresholding factor.

13. The method according to claim 12 , wherein determining a cluster of the thresholded slices includes: selecting a plurality of centrally located slices which correspond to the myocardium; determining a largest component within each selected slice according to connected component analysis; fitting a bounding box around each of the largest components; determining a longest continuous cluster of bounding boxes with adjacent centroids which differ by less than four pixels; and generating an initial ROI by masking the volume image representation with a largest bounding box within the determined cluster.

14. The method according to claim 13 , wherein cleaning the determined cluster of slices includes: thresholding each slice of the initial ROI according to a ratio of a maximum pixel intensity of each corresponding slice and a preselected thresholding factor; determining a centroid of each thresholded slice of the initial ROI; and removing slices from the thresholded ROI which exceed a predetermined Manhattan distance from the centroid of a central slice.

15. The method according to claim 10 , wherein isolating the left ventricle includes: selecting a central slice of the cleaned slices; determining a maximum and a minimum horizontal tangent point and a maximum and a minimum vertical tangent point; estimating a first cardiac angle of a line between the minimum vertical tangent point and the maximum horizontal tangent point; and estimating a second cardiac angle of a second line between the maximum vertical tangent point and the minimum horizontal tangent.

16. A functional imaging system, comprising: at least one detector head configured for acquiring functional imaging data; a controller configured to control the acquisition of the functional image data and/or a position of the detector head; a region of interest processor configured to determine a region of interest (ROI) encompassing at most a myocardium from the acquired functional imaging data; a parameterization processor configured to estimate and/or quantifies parameters from the determined ROI; and an electronic data processing component programmed to: generate projection image representations from the functional image data, threshold the projection image representation based on a histogram of a selected representative slice, determine a localized cardiac region within the thresholded data according to at least one of a horizontal ray casting profile and anatomical heuristics, determine extents of the localized cardiac region according to image based and heuristics based cleaning techniques, and reconstruct the determined image data which corresponds to the determined extents of the localized cardiac region.

17. The functional imaging system according to claim 16 , wherein the estimated diagnostic parameters include at least one of myocardial blood flow, regional myocardial blood flow, flow reserve, and ejection fraction.

18. A non-transitory computer readable storage medium containing instructions, when executed by a computer, causes the computer to carry out the method according to claim 1 .

19. A non-transitory computer readable storage medium containing instructions, when executed by a computer, causes the computer to carry out the method according to claim 10 .

20. The method according to claim 1 , wherein in the acquisition step the functional image data is acquired by a nuclear imaging system and the determination and estimation steps are performed with a computer processor and further comprising displaying the localized cardiac region on a display apparatus.

21. The method according to claim 1 wherein acquiring the functional image data includes acquiring nuclear image data from a radioisotope injected into the subject with one of a single photon emission computed tomography system (SPET).

22. An imaging system, comprising: at least one detector head configured for acquiring functional imaging data; a controller configured to control the acquisition of the functional image data and/or a position of the detector head; an electronic data processing component programmed for: reconstructing the functional imaging data into a volume image representation; thresholding transverse slices of the volume image representation according to a maximum pixel intensity of each corresponding slice; determining a cluster of the thresholded slices which correspond to a left ventricle; cleaning the determined cluster of slices according to a predetermined Manhattan distance between a left ventricle centroid in a central slice to the left ventricle centroid in the remaining clustered slices; and isolating the left ventricle in the cleaned slices according to cardiac angles of a septum of the myocardium; and a display apparatus configured to display the isolated left ventricle.